Immunocompetent mouse models revealed that S100A4⁺ monocytes/macrophages facilitate long-term Zika virus infection in the testes

ABSTRACT

During its global epidemic, Zika virus (ZIKV) attracted widespread attention due to its link with various severe neurological symptoms and potential harm to male fertility. However, the understanding of how ZIKV invades and persists in the male reproductive system is limited due to the lack of immunocompetent small animal models. In this study, immunocompetent murine models were generated by using anti-IFNAR antibody blocked C57BL/6 male mice and human STAT2 (hSTAT2) knock in (KI) male mice. After infection, viral RNA could persist in the testes even after the disappearance of viremia. We also found a population of ZIKV-susceptible S100A4⁺ monocytes/macrophages that were recruited into testes from peripheral blood and played a crucial role for ZIKV infection in the testis. By using single-cell RNA sequencing, we also proved that S100A4⁺ monocytes/macrophages had a great impact on the microenvironment of ZIKV-infected testes, thus promoting ZIKV-induced testicular lesions. In conclusion, this study proposed a novel mechanism of long-term ZIKV infection in the male reproductive system.

KEYWORDS:

• ZIKV
• mouse models
• human STAT2 KI mice
• S100A4
• testis

Introduction

During its epidemic in nearly 100 countries and territories, Zika virus (ZIKV) caused multiple nervous system syndromes, including microcephaly in the fetuses of infected mothers and Guillain‒Barré syndrome in adults. After studying many ZIKV-infected patients, it was found that ZIKV is sexually transmissible with male reproductive tract tropism, and most sexually transmitted cases were from males to females [1]. The first patient who acquired ZIKV through sexual contact was infected by her husband who returned to the United States from Senegal, where he had been infected [2]. After its epidemic in South America, ZIKV was then detected in the semen of 5/14 male patients, even after viremia disappeared [3], indicating a potential transmission of ZIKV through semen. Moreover, the typical reproductive system symptoms, including hematospermia and prostatitis, and decreasing sperm count of male patients also suggested a potential harm of ZIKV to male reproductive health [2,4,5]. Therefore, revealing the mechanism underlying ZIKV shedding in the male reproductive tract is of great importance to prevent sexual transmission and potential injury to the reproductive health of male patients.

The testes are vital organs for producing sperm and maintaining male fertility, and it has been proven that they are the specific target organs of ZIKV [6,7]. Monocytes/macrophages, including testicular macrophages, are susceptible to ZIKV [8-10]. Our previous research found that a subpopulation of myeloid-derived S100A4⁺ monocytes/macrophages was capable of supporting ZIKV replication and was recruited into the testes of IFNAR-deficient A6 mice during infection [11]. However, unlike human STAT2 (hSTAT2), murine STAT2 cannot be degraded by ZIKV NS5, and most studies have to be performed using immunodeficient murine models (such as IFNAR-deficient A6 mice and IFN-alpha/beta and IFN-gamma receptor-deficient AG6 mice), that produce more severe testicular lesions than those of male patients. Therefore, immunocompetent murine models, compared immunocompromised murine models, will improve the understanding of the testicular pathogenesis and immune response of ZIKV patients.

In this study, two immunocompetent murine models were used to investigate the mechanism of ZIKV infection in testes: one was generated by administering anti-IFNAR antibodies to C57BL/6 male mice one day before infection (Ab-blocked mice) [12,13], and the other was hSTAT2 knock in (KI) male mice [14]. We found that these models exhibited reversible testicular injury after infection. Viral RNA could still persist in the testes even after viremia disappeared. Single-cell RNA sequencing (scRNA-Seq) further proved that monocytes/macrophages promoted ZIKV-induced testicular lesions by participating in the classical activation of the complement system. Moreover, ZIKV-infected S100A4⁺ monocytes/macrophages persisted in the testes even after clearance in peripheral blood. Therefore, our study suggested that S100A4⁺ monocytes/macrophages were reservoirs of ZIKV, which promoted long-term infection and injury in ZIKV-infected testes.

Results

Immunocompetent ZIKV murine model exhibited mild and self-limiting symptoms

To evaluate ZIKV-induced testicular pathogenesis in immunocompetent models, mice were intraperitoneally (i.p.) infected with 1 × 10⁴ PFU ZIKV, and C57BL/6 male mice were i.p. administered with an anti-IFNAR antibody one day before infection (Ab-blocked mice) [12,13]. Mice injected with PBS served as controls. Partial ZIKV-infected mice exhibited mild symptoms, including ruffled hair or the hunchbacked appearance (Figure 1A & B), and no significant difference was observed between body weights of ZIKV-infected mice and controls (Figure 1A & B). All ZIKV-infected mice survived during the observation period (Figure 1A & B). These results indicated that the symptoms of ZIKV-infected immunocompetent male mice were mild and self-limiting.

Figure 1. Incidence of ZIKV-infected immunocompetent murine models.(A & B) Body weights, symptom scores and survival rates of ZIKV-infected Ab-blocked mice (A) or hSTAT2 KI mice (B) were monitored daily (n = 3–4 mice for each group);(C & D) ZIKV viral load and detection rate in peripheral blood of ZIKV-infected Ab-blocked mice (A) or hSTAT2 KI mice (B) were detected by RT-qPCR (n = 3–7 mice for each group);(E) Representative graph of plaque assay of serum from ZIKV-infected Ab-blocked mice and hSTAT2 KI mice;(F) ZIKV viral detection rate in major organs of immunocompetent murine models (n = 3-6 mice for each group). Results were shown as means ± SEM and analyzed using the two-sided Student’s t test. *p < 0.05, **p < 0.01.

Viral load in tissues of ZIKV-infected mice was analyzed by RT-qPCR. Viremia of Ab-blocked mice disappeared at 10 dpi (Figure 1C). Viremia of ZIKV-infected hSTAT2 KI mice significantly decreased from 3 to 21 dpi, viremia could still persist in all hSTAT2 KI mice during infection (Figure 1D). A plaque assay also confirmed a decreasing trend of viral load in serum (Figure 1E). In contrast, no viral RNA was detected in the liver, lung and kidney tissues from mice in either infection model (Figure 1F).

These results indicated that mice in both immunocompetent models were susceptible to ZIKV and exhibited self-limiting symptoms. More importantly, contrary to the immunodeficient mouse models who hold a persistent viremia as reported, immunocompetent models could gradually clear viruses in the serum, and therefore, they are more suitable to investigate ZIKV infection.

ZIKV RNA persists in the testes after viremia disappearance

As the testes are the specific target organs of ZIKV, we analyzed testicular pathogenesis in these immunocompetent murine models. Compared with those from controls, the testes from ZIKV-infected mice showed decreased size and weight (Figure 2A, B & G, H). However, there was no significant different in testicular weight at 7–21 dpi (Figure 2B & H). Spermatogenic cells were arranged more loosely, and their number was slightly reduced in the testes from Ab-blocked mice compared to control mice from 7 to 14 dpi (Figure 2C & I). Compared to ZIKV-infected Ab-blocked mice, hSTAT2 KI mice demonstrated less testicular pathogenesis at 7–21 dpi, which was demonstrated by less interstitial inflammatory cells and slighter seminiferous tubule injury (Figure 2C & I).

Figure 2. Testicular pathogenesis of ZIKV-infected immunocompetent murine models.(A & B) Representative pictures (A) and weight (B, n = 4-6 testes for each group) of testes from ZIKV-infected Ab-blocked mice; (C) H&E staining of testicular sections from ZIKV-infected Ab-blocked mice at 0-14 dpi. Scale bar, 25 μm;(D-F) ZIKV viral load (D) and detection rate (E) in testes of ZIKV-infected Ab-blocked mice were detected by RT-qPCR (n = 4 testes for each group). Distribution of ZIKV antigens were analyzed by IFA (F). Scale bar, 25 μm;(G & H) Representative pictures (G) and weight (H, n = 4-6 testes for each group) of testes from ZIKV-infected hSTAT2 KI mice; (I) HE staining of testicular sections from ZIKV-infected hSTAT2 KI mice at 0–21 dpi. Scale bar, 25 μm;(J–L) ZIKV viral load (J) and detection rate (K) in testes of ZIKV-infected hSTAT2 KI mice were detected by RT-qPCR (n = 6 testes for each group). Distribution of ZIKV antigens were analyzed by IFA (L). Scale bar, 25 μm. Results were shown as means ± SEM and analyzed using the two-sided Student’s t test. *p < 0.05, **p < 0.01. Nuclei were stained with DAPI.

To investigate ZIKV dynamics in testes, we analyzed the testicular viral load by RT-qPCR. In Ab-blocked mice, viral RNA was detected in 50% of testes at 7 dpi. After viremia disappeared, viral RNA was still detected in testicular tissues of some mice at 14 dpi (Figure 2D, E). Contrary to the decreasing trend of viral load in peripheral blood, viral load in the testes from ZIKV-infected hSTAT2 KI mice slightly increased from 7 to 21 dpi (Figure 2I, J). Viral antigens were detected in ZIKV RNA positive testes, and as revealed by immunofluorescence staining (IFA), most of them were distributed in the interstitium (Figure 2F & L).

The blood-testis barrier (BTB), one of the tightest barriers in mammalian tissues and is mainly composed of tight junction proteins. The BTB separates spermatogenic cells from interstitial immune cells, thus protecting spermatogenic cells from immune attack and pathogen invasion. We further analyzed the distribution of tight junction proteins in ZIKV-infected testes. IFA showed that ZO-1 and Occludin were distributed in a continuous linear pattern along the outer edge of the seminiferous tubules in control testes. After infection, Occludin was significantly ldownregulated and discontinuously distributed in both models (Figure 3A, D). ZO-1 was also slightly downregulated (Figure 3A, D). The expression of tight junction proteins was quantified by Image J. We found that the expression of Occludin and ZO-1 slightly decreased in infected testes compared to controls (Figure 3B, C and E, F). Interestingly, these tight junction proteins recovered their expression at 21 dpi, (Figure 3B, C and E, F), suggesting the testicular injury was reversible in ZIKV-infected immunocompetent murine models.

Figure 3. Distribution of tight junction proteins in testes of ZIKV-infected immunocompetent murine models.(A–C) Distribution of Occludin (A) from testes from ZIKV-infected immunocompetent murine models was detected by IFA. Scale bar, 25 μm. The percentage area of Occludin% from ZIKV-infected Ab-blocked mice (B) and hSTAT2 KI mice (C) were analyzed by Image J, (n = 4 testes for each group);(D–F) Distribution of ZO-1 (D) from testes from ZIKV-infected immunocompetent murine models was detected by IFA. Scale bar, 25 μm. The percentage area of ZO-1% from ZIKV-infected Ab-blocked mice (E) and hSTAT2 KI mice (F) were analyzed by Image J, (n = 4 testes for each group). Results were shown as means ± SEM and analyzed using the two-sided Student’s t test. *p < 0.05, **p < 0.01. Nuclei were stained with DAPI.

S100A4⁺ monocyte/macrophage infiltration assisted the expression of proinflammatory cytokines

The testes are well-known immune immune-privileged organs, and the infiltration of inflammatory cells in the interstitium affects their physiological function. Therefore, we analyzed inflammatory cell infiltration in the ZIKV-infected testes. We found that most of the inflammatory cells were S100A4⁺ monocytes/macrophages, while T cells were rarely observed (Figure 4A). In addition, no positive stains of ZIKV antigens were observed in CD8⁺ or CD4⁺ T cells (Figure S1A, B).

Figure 4. S100A4⁺ monocytes/macrophages infiltration leads to persistent high expression of pro-inflammatory cytokines. (A) Distribution of S100A4⁺ cells in testes of ZIKV-infected murine models were detected by IFA. Scale bar, 25 μm;(B–F) The expression levels of major proinflammatory cytokines (B-E) and number of S100A4⁺ cells (F) in testes from ZIKV-infected Ab-blocked mice (n = 4–5 testes for each group). (G–K) The expression levels of major proinflammatory cytokines (G-J) and number of S100A4⁺ cells (K) in testes from ZIKV-infected hSTAT2 KI mice (n = 4–6 testes for each group). Results were shown as means ± SEM and analyzed using the two-sided Student’s t test. *p < 0.05, **p < 0.01. Nuclei were stained with DAPI.

Proinflammatory cytokines, including IL-1β, IL-6, IL-10 and TNF-α were quantified by RT-qPCR. IL-1β and TNF-α were significantly upregulated after infection (Figure 4B–E & G–J). IL-1β, which was mainly derived from monocytes/macrophages, reached its peak level at 7 dpi. then slightly decreased (Figure 4F & K). This trend was consistent with the number of S100A4⁺ monocytes/macrophages. Therefore, we further analyzed the expression of IL-1β in S100A4⁺ monocytes/macrophages and found significant costaining of S100A4 with IL-1 in ZIKV-infected testes (Figure S2). These results indicated that S100A4⁺ monocytes/macrophages could promote ZIKV-induced testicular inflammation by secreting IL-1β.

The permeability of the BTB is regulated by a variety of cytokines, including TNF-α. Therefore, the upregulated expression of TNF-α might cause BTB impairment. To investigate the effect of TNF-α on tight junction proteins in vitro, Sertoli cells were treated with TNF-α before analyzed with IFA. ZIKV-infected or uninfected Sertoli cells served as infected control or uninfected control cells respectively. In uninfected cells, ZO-1 and Occludin were mainly distributed around the boundary of Sertoli cells in continuous line. Consistent with the in vivo results, ZO-1-positive staining significantly decreased after treatment for 24 and 48 h (Figure 5A, C). The expression level of Occludin showed no obvious changes at 24 h but significantly decreased at 48 h (Figure 5B, D).

Figure 5. Effect of ZIKV infection and TNF-α on the distribution of ZO-1 and Occludin in vitro. Sertoli cells were infected by ZIKV (multiplicity of infection, MOI = 1) or treated with 50 ng TNF-α at 32˚C and were collected at 24 and 48 h. Distribution of ZO-1 (A & B) and Occludin (C & D) were analyzed by immunofluorescent staining. Nuclei were stained with DAPI. Scale bar, 25 μm.

To further confirm whether S100A4⁺ monocyte/macrophage infiltration promoted ZIKV-induced testicular lesions, ZIKV-infected hSTAT2 mice were intraperitoneally injected with niclosamide, an FDA-approved medicine that has been reported as an S100A4 inhibitor [15]. The weight and size of the testes were significantly increased in niclosamide-treated hSTAT2 KI mice compared to mock-treated mice (Figure 6A, B). HE staining also revealed less severe seminiferous tubule morphology in niclosamide-treated mice (Figure 6C). Moreover, the testicular viral load of niclosamide-treated mice was significantly lower than that of mock-treated mice (Figure 6D). We found that niclosamide treatment significantly blocked the infiltration of most S100A4⁺ monocytes/macrophages in ZIKV-infected testes (Figure 6C). In line with this finding, the expression levels of proinflammatory cytokines in the testes of niclosamide-treated mice were also lower than those in the testes of mock-treated mice (Figure 6E–I).

Figure 6. Niclosamide alleviated ZIKV-induced testicular injury in hSTAT2 KI mice.(A and B) The representative pictures (A) and weight (B) of testes at 14 dpi. (n = 6-8 testes for each group); (C) HE staining and IFA of testes at 14 dpi. Testicular sections were co-immunostaining with anti-ZIKV antigens antibody and anti-S100A4 antibody. Scale bar, 25 μm;(D) ZIKV viral load of testes at 14 dpi. (n = 6 testes for each group); (E–I) The expression levels of proinflammatory cytokines in testes at 14 dpi. (n = 4–6 testes for each group). Results were shown as means ± SEM and analyzed using the two-sided Student’s t test. *p < 0.05, **p < 0.01. Nuclei were stained with DAPI.

All these results suggested that S100A4⁺ monocytes/macrophages promote ZIKV-induced testicular lesions and inflammation.

Single-cell RNA sequencing revealed reduced monocyte/macrophage infiltration in ZIKV-infected testes from hSTAT2 KI mice

To characterize the roles of S100A4⁺ monocytes/macrophages in ZIKV-induced testicular lesions, testicular cells from ZIKV-infected hSTAT2 mice (14 dpi.) were subjected to scRNA-Seq. Cells from PBS-injected mice served as controls.

After filtering out poor-quality cells, 12,455 cells in control testes and 10,110 cells in ZIKV-infected testes were selected (Figure 7A, B and Figure S3A, B). These cells were grouped into 5 cell clusters (Figure 7A, B and Figure S3A, B) based on the shared nearest neighbour (SNN) module optimization algorithm and the expression of known cell type-specific markers [16]. The number of spermatogenic cells was significantly decreased after infection (Figure 7A, B and Figure S3A, B), and several cell death-related genes were significantly upregulated in the spermatogenic cells from ZIKV-infected testes (Figure 7C). Moreover, the number of immune cells/Leydig cells was slightly increased (Figure 7A, B and Figure S3A, B). We then identified and analyzed the types of immune cells and found that the number of monocytes/macrophages was higher than that of other immune cells (T/NK cells, B cells) (Figure 7D and Figure S3C).

Figure 7. Characterization of cells in control and ZIKV-infected testes from hSTAT2 KI mice analyzed by Single-cell sequencing.(A) The distribution of cell clusters was shown in UMAP chart;(B) The number of cell clusters in the testes from control and ZIKV-infected mice; (C) Expression of genes related to apoptosis, pyroptosis, necrosis, autophagy, and ferroptosis in spermatogenic cells in the testes from control and ZIKV-infected mice; (D) The distribution of cell clusters in Immune cells/Leydig cells was shown in UMAP chart;(E) The expression of genes involved in classical complement activation pathway in all cell clusters;(F–H) The expression of genes involved in classical complement activation pathway in monocytes and macrophages were shown in bubble chart (F), heatmap (G) and violin chart (H);(I and J) The expression level of C1q in ZIKV-infected testes from hSTAT2 KI mice were analyzed by RT-qPCR (I) and ELISA (J) ZIKV-infected mice. (n = 4 testes for each group);(K and L) The number (K) and distribution (L) of C1q⁺ cells in ZIKV-infected testes from hSTAT2 KI mice at 14 dpi. Scale bar, 25 μm;(M) The co-immunostaining of DDX4 and MAC in testes from ZIKV-infected testes of hSTAT2 KI mice. Scale bar, 25 μm; Results were shown as means ± SEM and analyzed using the two-sided Student’s t test. *p < 0.05, **p < 0.01. Nuclei were stained with DAPI.

Our previous study showed that classical activation of the complement system in ZIKV-infected testes was the main cause of ZIKV-induced spermatogenic cell damage [17]. In this study, scRNA-Seq with ZIKV-infected testes of hSTAT2 KI mice also showed that the expression level of C1q, the key molecule of classical complement activation, was upregulated upon ZIKV infection and its main source were monocytes/macrophages (Figure 7E–H). These results suggested that monocytes and macrophages promoted the classical complement activation in ZIKV-infected testes from immunocompetent hSTAT2 KI mice.

Interestingly, compared with our previous scRNA-Seq results in immunocompromised A6 mice [17], the number of infiltrated monocytes/macrophages was significantly lower. We hypothesized that lower monocyte/macrophage infiltration might cause slightly complement activation in ZIKV-infected testes from hSTAT2 KI mice. ELISA and RT-qPCR showed that although the expression level of C1q was higher in ZIKV-infected testes than in the controls, it was not significantly increased from 7-21 dpi (Figure 7I, J). More importantly, the number of C1q⁺ S100A4⁺ cells in testes from niclosamide-treated ZIKV-infected hSTAT2 KI mice was even lower than that in mock-treated mice (Figure 7K, L). Less costaining of the membrane attack complex (MAC), the cytolytic effector of the complement system [18], with spermatogenic cells was also detected in the testes from niclosamide-treated hSTAT2 KI mice compared with mock-treated mice (Figure 7M). Therefore, these results indicated that the lower monocyte/macrophage infiltration in ZIKV-infected testes from immunocompetent hSTAT2 KI mice caused mild testicular injury.

S100A4⁺ monocytes/macrophages prolong ZIKV infection in the testes

Monocytes/macrophages are susceptible to several flaviviruses [19-24]. Our previous studies proved that myeloid-derived S100A4⁺ monocytes/macrophages could promote ZIKV infection in the testes from immunodeficient A6 mice. In this study, we further explored whether S100A4⁺ monocytes/macrophages promoted long-term ZIKV infection in the testes of immunocompetent mice.

First, the expression level of the S100a4 gene in ZIKV-infected testes was analyzed by RT‒qPCR. S100a4 expression was significantly upregulated at 7 dpi in the testes from Ab-blocked mice compared with the testes from control mice. At 14 dpi, when ZIKV RNA was only detected in one testicular sample (1/6) from ZIKV-infected Ab-blocked mice (Figure 2D), S100a4 expression in viral RNA-negative testes reduce to the same level in control testes, while that in viral RNA-positive testes was still higher than all control testes (Figure 8A, showed by arrow). In contrast, S100a4 expression in the testes of ZIKV-infected hSTAT2 KI mice was significantly upregulated from 7 to 21 dpi (Figure 8B). IFA revealed that S100A4⁺ monocytes/macrophages have infiltrated in the interstitium and were costained with ZIKV antigens (Figure 8C). Notably, in the testes from Ab-blocked mice with viral antigens detected at 14 dpi, most ZIKV antigens also colocalized with S100A4⁺ monocytes/macrophages (Figure 8C). Therefore, S100A4⁺ monocytes/macrophages could assist the long-term infection of ZIKV in the testes.

Figure 8. S100A4⁺ monocytes/macrophages prolong the presence of ZIKV infection in testes.(A & B) Expression level of S100a4 gene in testes from ZIKV-infected Ab-blocked mice (A) and hSTAT2 KI mice (B) were detected by RT-qPCR (n = 4-6 testes for each group);(C) Distribution of ZIKV antigens and S100A4 in testes from ZIKV-infected murine models were analyzed by IFA. Scale bar, 25 μm;(D) ZIKV-infected S100A4⁺ monocytes in peripheral blood from ZIKV-infected Ab-blocked mice at 0–14 dpi by were detected flow cytometry (n = 3–4 mice for each group);(E-H) Expression levels of Ccl2 and Ccl5 in testes from ZIKV-infected Ab-blocked mice (E, F) and hSTAT2 KI mice (G, H) were detected by RT-qPCR (n = 3–6 testes for each group). Results were shown as means ± SEM and analyzed using the two-sided Student’s t test. *p < 0.05, **p < 0.01. Nuclei were stained with DAPI.

Second, as monocytes are the main target cells of ZIKV in peripheral blood [25], we then analyzed the relationship between S100A4⁺ monocytes and the duration of viremia. ZIKV-infected S100A4⁺ monocytes in the peripheral blood from Ab-blocked mice at 0–14 dpi. were detected by flow cytometry. The quantity of S100A4⁺ CD11b⁺ monocytes was significant decreasing in peripheral blood at 7 dpi followed by a significantly decrease trend at 14 dpi (Figure 8D). We also found that co-staining of S100A4⁺ CD11b⁺ cells and ZIKV antigens at 7 dpi and they disappeared at 14 dpi (Figure 8D). Therefore, these results indicated a consistency of the clearance of ZIKV-infected S100A4⁺ monocytes with the disappearance of viremia (Figure 1C).

Finally, as the recruitment of monocytes/macrophages into specific tissues depends on the high expression of chemokines, we hypothesize that the persistence of S100A4⁺ monocytes/macrophages in ZIKV-infected testes might be caused by the high expression of monocyte/macrophage recruitment-related chemokines. RT-qPCR showed that the expression of CCLs (including monocyte/macrophage recruitment related chemokines CCL2 and CCL5) was significantly upregulated from 7 to 14 dpi (Figure 8E–H), indicating that ZIKV might promote the recruitment of peripheral blood-derived S100A4⁺ monocytes into testes by upregulating the expression of CCLs, thus prolonging the presence of ZIKV in testes.

These results suggested that S100A4⁺ monocytes/macrophages were the main target cells of ZIKV, and could assist the long-term infection of ZIKV in testes.

Discussion

ZIKV has attracted worldwide attention due to its close relationship with severe neurological symptoms, including Guillain‒Barré syndrome and microcephaly [26-29]. The major significant difference between ZIKV and other mosquito-borne flaviviruses is that ZIKV can be transmitted sexually and persist in the male reproductive system. Although researchers have revealed the possible mechanism of ZIKV-induced neurological lesions, the mechanism underlying ZIKV invasion and persistence in the male reproductive system is still hindered by a lack of data from immunocompetent murine models. In this study, by using two immunocompetent murine models, we demonstrated that S100A4⁺ monocytes/macrophages were susceptible to ZIKV and were recruited into the testes during ZIKV infection, thereby prolonging ZIKV infection in the testes after the disappearance of viremia. Moreover, by using scRNA-Seq, we also proved that S100A4⁺ monocytes/macrophages the damaged testicular microenvironment by widely communicating with immune cells and testicular resident cells, thus promoting the activation of the proinflammatory response in ZIKV-infected testes.

ZIKV has been proven to be involved in hemospermia, microspermia and prostatitis [2,4,30-32]. Researchers also reported a decreasing trend of sperm count in male patients from 119 × 10⁶ at 7 dpi. to 45.2 × 10⁶ at 30 dpi [33]. More importantly, the infectious virus particles and RNA could persist in semen for 69 and 188 days respectively [34-38]. The virus could still be detected in the semen of some male patients even after viremia disappeared [34,35]. By using immunodeficient murine models, researchers proved that the testes are the specific target organs of ZIKV and that long-term ZIKV infection can occur in the testes [39-43]. However, since immunodeficient mice demonstrated severe testicular lesions and continuous ZIKV viremia [11], these mice could not fully reproduce the testicular pathology and viral kinetics of male patients. To better reproduce the clinical symptoms of male patients, two immunocompetent murine models were used in this study. As ZIKV failed to antagonize Stat2-dependent interferon (IFN) responses in mice [44-46], C57BL/6 mice were intraperitoneally injected with anti-IFNAR antibody and intraperitoneally infected with ZIKV the next day. hSTAT2 KI mice on a C57BL/6 genetic background, which were reported to be susceptible to mouse-adapted ZIKV (ZIKV-Dak-MA) [14], were also intraperitoneally infected. These mice exhibited milder symptoms and viremia clearance, which suggested that the viral kinetics of these infection models are similar to those of patients.

Tropism to the male reproductive tracts is one of the major significant differences between ZIKV and other mosquito-borne flaviviruses, and testicular injury is also a major difference [11,17,39-41]. We then analyzed ZIKV-induced testicular injury in immunocompetent murine models. The size and weight of infected testes were significantly reduced at 7 dpi. However, at 14-21 dpi, contrary to the continuous aggravation of testicular lesions in immunodeficient mice, testicular atrophy did not progress in immunocompetent mice. IFA and WB further revealed the downregulated and discontinuous distribution pattern of tight junction proteins in ZIKV-infected testes compared with control testes at 7 dpi and the high expression at 14–21 dpi. compared with 7 dpi, thus suggesting a distinct pathological process of ZIKV-infected immunocompetent murine testes. More importantly, viral dynamics in the testes of immunocompetent mice were similar to those in the semen from male patients, as ZIKV could persist in the testes after viremia vanished. Therefore, these models are suitable for investigating the mechanism of ZIKV persistence in the testes.

Monocytes/macrophages have been proven to be susceptible to flaviviruses. CD206⁺ macrophages/moDCs are important targets for visceral ZIKV replication following hematogenous dissemination of ZIKV from the infected site [9]. Tissue macrophages can also promote ZIKV infection in numerous organs, including testicular macrophages in the testes and Hofbauer cells in the placenta [9,10,47]. Our previous study also found that a population of ZIKV susceptible myeloid-derived S100A4⁺ monocytes/macrophages could promote the long-term infection of ZIKV in the testes from immunocompetent A6 mice [9]. In this study, massive infiltration of ZIKV-infected S100A4⁺ monocytes/macrophages in testicular interstitium of immunocompetent murine models was also observed. We also proved that these cells were the main infiltrating immune cells in ZIKV-infected testes from immunocompetent mice. The infiltrated inflammatory cells also suggested a potential inflammatory process in ZIKV-infected testes. Abnormally high levels of proinflammatory cytokines in semen (IL-1β, IL-6, TNF-α, etc.) have proven to be associated with decreased sperm motility [48,49]. Inflammatory cytokines also exacerbate oxidative stress, thereby compromising sperm integrity [50]. In this study, the expression levels of various proinflammatory cytokines (IL-1β, IL-6, IL-10 and TNF-α) in the testes of immunocompetent murine models were increased after infection. IFA also confirmed the colocalization of S100A4⁺ monocytes/macrophages with IL-1β. Moreover, by using scRNA-Seq, we confirmed that S100A4⁺ exogenous monocytes/macrophages were the main infiltrating cells in ZIKV-infected testes from immunocompetent mice. scRNA-Seq further indicated that these cells can promote ZIKV-induced testicular inflammation, they secrete secret proinflammatory cytokines and participate in the activation of proinflammatory pathways. Therefore, S100A4⁺ monocytes/macrophages promoted the inflammatory response of ZIKV-infected testes.

Monocyte/macrophage migration is promoted by monocyte/macrophage chemotactic factors (MCFs) [51-53]. Our results proved that ZIKV infection promoted the recruitment of S100A4⁺ monocytes/macrophages into the testes by upregulating the expression of CCL2 and CCL5. In addition, the presence of ZIKV-infected S100A4⁺ monocytes in peripheral blood was consistent with the duration of viremia. Furthermore, ZIKV-infected S100A4 + monocytes/macrophages could still be detected in testes after viremia disappeared, thus suggesting that ZIKV-infected monocytes were one of the main causes that promoted ZIKV viremia. ZIKV infection in the testes may be prolonged by the recruitment of ZIKV-infected S100A4⁺ monocytes from peripheral blood.

In summary, our study demonstrated that monocytes/macrophages could promote ZIKV infection in immunocompetent murine models, and proved that ZIKV-infected S100A4⁺ monocytes/macrophages were recruited into ZIKV-infected testes by CCLs, thus prolonging ZIKV infection in the testis after viremia disappeared. Our study revealed a new mechanism by which ZIKV establishes long-term infection in testes and provided a potential theory for protecting male fertility in male ZIKV patients.

Material and methods

Virus

Asian ZIKV (strain SMGC-1, GenBank accession number: KX266255) was propagated in C6/36 Aedes albopictus cells. Plaque assay on Vero cells with 1.0% methylcellulose was subjected to determine viral titer. Virus was stored at −80°C until use.

Mouse experiments

SPF C57BL/6 mice were purchased from Charles River company (Beijing) by the Animal Department of Capital Medical University, bred under specific pathogen-free conditions at Capital Medical University. hSTAT2 KI mice were purchased from the Jackson Laboratory. 6–8 weeks old male C57BL/6 male mice were intraperitoneally injected (i.p.) with anti-IFNAR antibody (1 mg per mice) one day before intraperitoneally infected with 1 × 10⁴ PFU ZIKV. 6–8 weeks old male hSTAT2 KI mice were also intraperitoneally infected with 1 × 10⁴ PFU ZIKV. Testes were harvested and immediately fixed in Modified Davidson’s Fluid solution (5 mL of glacial acetic acid, 15 mL of ethanol, 30 mL of 40% formaldehyde and 50 mL of distilled water) overnight before dehydrated and paraffin-embedded.

Plaque assay

After centrifugation, serum from ZIKV-infected mice was transferred to a Vero cell monolayer in 24-well plate. Control wells were incubated with serum from uninfected mice. After 1 h for absorption at 37 °C, 5%CO₂ with gentle rocking every 15 min, cells were then washed by MEM containing 2% FBS and overlaid with MEM containing 2% FBS and 1.2% methylcellulose. After incubated at 37 °C, 5%CO₂ for 7days, cells were stained by saturated crystal violet solution to visualize the plaques.

Testicular sections (5 µm in thickness) were conventional dewaxed to water before staining with hematoxylin for 12 min. After stained with eosin for 20 min, sections were dehydrated by re-immersing in alcohol and xylene, then mounted using synthetic resin.

Immunofluorescence staining (IFA)

Testicular sections were incubated with the following primary antibodies overnight at 4 °C after conventional dewaxed to water, including rabbit anti-mouse rabbit anti-mouse CD8α (1:400, Cell Signalling Technology, D4W2Z), rabbit anti-mouse CD4 (1:200, Abcam, ab183685) Occludin (1:200, Abcam, ab167161), rabbit anti-mouse ZO-1 (1:200, Thermo Fisher, 33-9100), rabbit anti-mouse S100A4 (1:500, Cell Signaling Technology, 13018S), mouse anti-mouse S100A4 (1:500, proteintech, 66489–1), rabbit anti-mouse IL1-beta (1:500, Abcam, ab283818) or mouse anti-ZIKV antibody 4G2 (1:500). After washed, sections were incubated with secondary antibodies at 37 °C for 1 h in the dark condition, including donkey anti-mouse IgG (1:1000, Alexa Fluor R 488, A21202, Life technologies), donkey anti-rabbit IgG (1:1000, Alexa Fluor R 594, A21207, Life technologies). Images were then captured with Olympus microscope (IX71, Olympus, Japan).

ZIKV mRNA quantification and proinflammatory cytokines mRNA, monocyte/macrophage chemotactic factors mRNA and S100a4 mRNA relative quantification

Testes from ZIKV-infected and control mice were harvested at different time points as indicated. Samples were homogenated and was RNA extracted in TransZol (TransGen China, ET101-01) according the manufacturer protocol. Real-time qPCR was performed as previously reported with Quant One Step qRT-PCR (Tiangen, China) on 7500 Real Time PCR System (Applied Biosystems, USA). Copies of ZIKV mRNA were quantified by standard curve method. ZIKV genome RNA transcribed in vitro was quantified and used as a standard template to establish the standard curve. Relative expression of proinflammatory cytokines, MCFs and S100a4 was determined by using GAPDH as internal control and 2−ΔΔCt method. The primer sequences were as follows:

Table

ZIKV forward	5′-TCAGACTGCGACAGTTCGAGT-3′
ZIKV reverse	5′-GCATATTGACAATCCGGAAT-3′;
GAPDH forward	5′-GCATTGTGGAAGGGCTCA-3′;
GAPDH reverse	5′-ACCAGTGGATG CAGGGAT-3′
IL-1β forward	5′-GCAACTGTTCCTGAACTCAACT-3′
IL-1β reverse	5′-ATCTTTTGGGGTCCGTCAACT-3′[54]
IL-6 forward	5′-ACTTCCATCCAGTTGCCTTCTTGG-3′;
IL-6 reverse	5′-TTAAGCCTCCGACTTGTGAAGTGG-3′[55];
IL-10 forward	5′-TGCTATGCTGCCTGCTCTTA-3′
IL-10 reverse	5′-TCATTTCCGATAAGGCTTGG-3′[54]
TNFa forward	5′-CCCACTCTGACCCCTTTACT-3′;
TNFa reverse	5′-TTTGAGTCCTTGATGGTGGT-3′[56];
CCL2 forward	5′-TTGTCACCAAGCTCAAGAGAGA-3′;
CCL2 reverse	5′-GAGGTGGTTGTGGAAAAGGTAG-3′[57];
CCL5 forward	5′-ATCTTGCAGTCGTGTTTGTCAC-3′;
CCL5 reverse	5′-GGAGTGGGAGTAGGGGATTACT-3′[57].

Flow cytometry

For flow cytometry, peripheral blood from IFNAR antibody blocked mice were subjected to whole blood red cell lysing reagent to obtained peripheral blood mononuclear cells. After fixation and permeabilization, the cells were then incubated with fluorochrome-conjugated antibodies to CD11b-APC (1:100, e-bioscience, 17-0112-81), rabbit anti-mouse S100A4 (1:500, Cell Signaling Technology, 13018S) and mouse anti-ZIKV antibody 4G2 (1:500). FITC-conjugated goat anti-mouse IgG antibody (1:1000, Termo Fisher Scientifc, F2761) and donkey anti-rabbit IgG PE secondary antibody (1:1000, Bioscience, 12-4739-81e) were used as secondary antibodies. Samples were processed on DxFLEX flow cytometer (Beckman Coulter, USA) and data was analyzed using CytExpert software (version 2.0). Compensation was performed by conjugated with fluorescent antibodies for each channel.

Tissue dissociation and preparation of single-cell suspensions

Testes from ZIKV-infected hSTAT2 KI mice were harvested at 14 dpi. Testes from uninfected mice served as controls. Tissues were dissociated into single cells in dissociation solution (0.35% collagenase IV5, 2 mg/ml papain, 120 Units/ml DNase I) in 37 °C water bath with shaking for 20 min at 100 rpm. Digestion was terminated with 1× PBS containing 10% fetal bovine serum (FBS, V/V), then pipetting 5–10 times with a Pasteur pipette. The resulting cell suspension was filtered by passing through 70-30um stacked cell strainer and centrifuged at 300 g for 5 min at 4 °C. The cell pellet was resuspended in 100 μl 1× PBS (0.04% BSA) and added with 1 ml 1× red blood cell lysis buffer (MACS 130-094-183, 10×) and incubated at room temperature or on ice for 2-10 min to lyse remaining red blood cells. After incubation, the suspension was centrifuged at 300 g for 5 min at room temperature. The suspension was resuspended in 100 μl Dead Cell Removal MicroBeads (MACS 130-090-101) and remove dead cells using Miltenyi ® Dead Cell Removal Kit (MACS 130-090-101). Then the suspension was resuspended in 1× PBS (0.04% BSA) and centrifuged at 300 g for 3 min at 4 °C (repeat twice). The cell pellet was resuspended in 50 μl of 1×PBS (0.04% BSA). The overall cell viability was confirmed by trypan blue exclusion, which needed to be above 85%, single cell suspensions were counted using a haemocytometer/Countess II Automated Cell Counter and concentration adjusted to 700–1200 cells/μl.

Chromium 10× genomics library and sequencing

Single-cell suspensions were loaded to 10× Chromium to capture single cells according to the manufacturer’s instructions of 10× Genomics Chromium Single-Cell 3′ kit (V3). The following cDNA amplification and library construction steps were performed according to the standard protocol. Libraries were sequenced on an Illumina NovaSeq 6000 sequencing system (paired-end multiplexing run, 150 bp) by LC-Bio Technology co. ltd., (Hang Zhou, China).

Single-cell RNA-seq analysis

Sequencing results were demultiplexed and converted to FASTQ format using Illumina bcl2fastq software. Sample demultiplexing, barcode processing and single-cell 3′ gene counting by using the Cell Ranger v5.0.1 [58] and scRNA-seq data were aligned to Ensembl genome GRCm38 reference genome (release-95). The Cell Ranger output was loaded into Seurat v4.1.1 be used to Dimensional reduction, clustering, and analysis of scRNA-seq data [59]. Overall, 22,565 cells passed the quality control threshold: all genes expressed in less than one cells were removed, number of genes expressed per cell > 500 as low cut-off, the percent of mitochondrial-DNA derived gene-expression < 25%.

To visualize the data, we further reduced the dimensionality of all 22,565 cells using Seurat and used UMAP to project the cells into 2D space, The steps include: 1. Using the LogNormalize method of the “Normalization” function of the Seurat software to calculated the expression value of genes; 2. PCA (Principal component analysis) analysis was performed using the normalized expression value. Within all the PCs, the top 10 PCs were used to do clustering and t-SNE analysis; 3. To find clusters, selecting weighted Shared Nearest Neighbour (SNN) graph-based clustering method. Marker genes for each cluster were identified with the Bimod Likelihood-ratio test with default parameters via the FindAllMarkers function in Seurat. This selects markers genes which are expressed in more than 10% of the cells in a cluster and average log (Fold Change) of greater than 0.25. Expression of up-regulated and down-regulated genes analyzed by Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was conducted by cluster Profiler R package v4.2.2 [60].

Raw sequence data of single-cell RNA sequencing has been uploaded to SRA under BioProject PRJNA1028143.

Statistical analysis

Statistical analysis was performed by SPSS 17.0 Software (IBM, Armonk, NY, USA). Data between two groups with normal distributions was analyzed by the repeated-measures analysis of variance or the Student’s t test. All results were presented as the mean ± standard error of the mean (SEM) from at least three different repeats. P < 0.05 was considered as statistically significant between two groups.

Author contributions

W. Yang and C. Zhang performed experiments and analyzed the data. L.B. Liu and Z.Z. Bian performed animal experiments. J.T. Chang and D.Y. Fan helped to maintain cells, viruses, J. An, P.G. Wang and N. Gao conceived the project, analyzed the data, wrote and reviewed the manuscript. J. An coordinated the whole research.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

We appreciate the grants from the National Key Research and Development Plan of China (2021YFC2300202 to J.A.), The National Natural Science Foundation of China (U1902210 to J.A., 82372251 to P.G.W., 81871641 to P.G.W., 81972979 to J.A., and 82172266 to P.G.W.) and Natural Science Foundation of Beijing (7232002 to N.G.). The funders had no role in study design, data collection, and analysis, decision to publish, or preparation of the manuscript.